Zookeeper + Book Keeper seems to be a good candidate for this.
I've recently changed http://code.google.com/p/bigstreams from using
hazelcast to using zookeeper, it uses zookeeper to keep track of realtime
file transfers, and have found zookeeper to work great, failover is decent
and predictable also.
On Tue, Jul 19, 2011 at 9:22 PM, Thiago Veiga <[EMAIL PROTECTED]> wrote:
> Hi ,
> i would like to create a hadoop fail over system , mainly for the
> master node in my hadoop cluster.
> I can t use the linux HA so i have been trying another approach
> I thought zookeeper wold be a good option , but i am not sure about that
> Does someone have a suggestion ?